74 ALASKAN GLACIER STUDIES
be destroyed, with perhaps a larger area if the forward thrust reached further. It seemed probable that the latter would be the case, for the breaking observed in 1906 suggested that the maximum had not been reached. As a matter of fact the Atrevida Glacier had practically ceased its forward motion in August, 1906, and the termination of its advance, and its return to a condition of stagnation, have been almost as spectacular as was its transformation into a state of activity between August, 1905 and June, 1906.
In 1909 we reoccupied the sites from which the observations and photographs of 1906 and 1906 were made, and were, therefore, able to make exact comparisons of the condition in the several years (Pis. TZTt, and XXI, A). No new observations of Atrevida Glacier were made during the two weeks spent in Yakutat Bay in 1910. In 1909 it was clear that the glacier had advanced but little since last seen in August, 1906, that the wounds from the spasmodic advance had been to a great extent healed by ablation, that the glacier had again relapsed into a state of stagnation, and that it was now once more possible to travel over its surface. In 1909 the Atrevida Glacier was far more like what it was in 1905 than in 1906, but its surface was much rougher than in 1905. The greater part of the glacier was covered with a sheet of ablation moraine (PL XIX, B) and this extended much farther up the valley than it did in 1905. This is a remarkable fact, and no other explanation occurs to us than that the extension of moraine cover up the glacier valley is due to the downshaking of rock fragments from the steep valley walls during the 1899 earthquakes, thus furnishing an unusually great supply of debris that was above the snow line in 1905 but had been brought down into the zone of ablation by the rapid advance of 1906. A few small areas of clear ice appeared in the waste of moraine that covered the glacier above Terrace Point, most of them in the middle of the glacier, where one would naturally expect less debris from avalanches. This upper glacier surface was viewed and photographed at the same site on Terrace Point, in both 1906 and 1909 (PI. XIX) and the photographs show clearly a decided lowering of the ice surface in the interval of three years. It is impossible to state the exact amount, but it cannot be less than 50 feet and may be even more than 100 feet. This lowering of the ice may not all be the result of ablation, for there may have been a flattening through the forward creep of the lower ice of the glacier after the spasmodic advance ceased.
Between Amphitheatre Knob and Terrace Point, just where the valley glacier portion may be said to end, more detailed studies were made than elsewhere, because it was here that Russell crossed in 1890, and the junior author in 1905, while it was carefully examined from both the east and west sides in 1906. There was, therefore, a better basis for comparative study here than elsewhere. In 1909, therefore, this portion was examined and photographed from Amphitheatre Knob and from Terrace Point, as in 1906, and it was crossed, as in 1905. With the exception of a narrow medial area, in which small patches of clear ice appeared, the entire surface was covered with a sheet of ablation moraine as it was in 1905; but the surface was much rougher than then, the morainic hillocks being higher and the intervening pits and valleys deeper. No sign of the crevassed dome observed in 1905 was visible in 1909. As compared with 1906 the surface was of course far smoother, and when viewed from either side no crevasses were visible, and, excepting in the area mentioned, almost no ice could be seen. This contrasts strikingly with 1906 when the glacier was broken by a maze of crevasses, and more than half the surface was clear ice; and the contrast is brought out most strikingly by the photographs taken from the same site on Terrace Point in the two years. In these